Passive Acoustic Emission Sensing Principles

While conventional predictive maintenance relies exclusively on vibration sensors for rotating machinery, static vessels and pipelines do not vibrate until they are on the verge of rupture. In our AI platform, Acoustic Emission (AE) serves as our proprietary physical sensing advantage: whenever steel under operational stress undergoes micro-crack propagation, active dislocation movement, or corrosion oxide rupture, it abruptly releases stored elastic strain energy.

This energy radiates outwards through the structure as a high-frequency transient mechanical stress wave. This physical phenomenon is known as Acoustic Emission (AE).

Unlike traditional ultrasonic testing (UT), where a technician sends an active sound wave into the metal and measures the reflection, Acoustic Emission is passive listening. The asset itself generates the sound when it is under operational stress. If a crack isn't growing, there is no sound. But the moment a crack extends by even a fraction of a millimeter under pressure or thermal load, it broadcasts an acoustic signature.

Updem Monpod autonomous wireless acoustic emission hardware module with differential piezo sensors
THE MONPOD™: Autonomous wireless acoustic emission module with magnetic mount and dual differential piezoelectric transducers deployed in field testing.

Acoustic Emission vs. Vibration Analysis

Plant engineers frequently ask: "We already have vibration sensors on our plant pumps. Why can't we just stick vibration sensors on our storage tanks and reactors?"

The table below explains why vibration analysis is the wrong tool for static structures, and why Acoustic Emission is required.

Parameter Updem Acoustic Emission (AE) Traditional Vibration Analysis
Primary Asset Type Static Assets: Storage tanks, pressure vessels, reactors, steam headers, pipelines. Rotating Machinery: Pumps, motors, gearboxes, turbines, bearings.
Frequency Bandwidth 30 kHz to 750 kHz (Ultrasound). Far above human hearing and plant mechanical hum. 1 Hz to 10 kHz (Low frequency mechanical displacement and harmonics).
What It Detects Active crack growth: Micro-fractures, crystal slip, active pitting corrosion, and pressurized leaks. Mechanical motion: Unbalance, misalignment, bearing looseness, resonance.
What a Static Tank Does A static tank doesn't shake or vibrate. But its steel plates generate continuous AE bursts when cracking under hydrostatic load. A stationary tank produces zero measurable vibration until a catastrophic breach occurs.
Early Warning Window 14 to 30+ Days (Detects defect at the incubation and micro-fracture stage). 2 to 5 Days (Detects bearing spall once metal has already begun disintegrating).

How We Filter Background Plant Noise

Real industrial plants are noisy environments: heavy pumps rumble, steam hisses through valves, product flows create pipe friction, and rain impacts tank roofs. If an acoustic system triggered on every noise, maintenance teams would disable it within three days due to false alarm fatigue.

Updem solves this challenge through a multi-stage physical and digital filtering architecture:

STAGE 1: HARDWARE BANDPASS

Mechanical Resonant Tuning

Our piezoelectric crystal elements have mechanical resonance peaks tuned between 150 kHz and 300 kHz. Low-frequency plant rumble (sub-20 kHz) cannot mechanically excite the transducer crystal.

STAGE 2: EDGE WAVELET AI

Transient Burst Pattern Discrimination

Continuous fluid flow produces steady-state Gaussian white noise. In contrast, an authentic crack burst is an abrupt, high-energy transient spike with sharp rise times (< 5 µs) and exponential decay. Our edge processor separates them in real-time.

ASTM E1316 & ISO 12716 Acoustic Emission Burst Architecture

Standardized transient feature extraction: click any parameter below to inspect its physical definition and diagnostic significance.

RISE TIME (t_r) 6.2 µs
Time from first threshold crossing to peak amplitude. Steep rise (< 10 µs) separates true micro-cracking from mechanical sliding friction.
PEAK AMPLITUDE (A_max) 82.4 dBAE
Maximum voltage excursion referenced to 1 µV at sensor crystal. Indicates crack tip kinetic severity and acoustic energy release.
RINGDOWN COUNTS (N) 86 counts
Number of positive cycles crossing the trigger threshold. Correlates directly with flaw size and acoustic ringing duration.
BURST DURATION (D) 245 µs
Time from first threshold crossing to final threshold crossing before damping below the floating background noise floor.
MARSE ENERGY 1,420 eu
Measured Area of the Rectified Signal Envelope. Directly proportional to physical elastic strain energy released during localized material fracture.
FLOATING THRESHOLD (V_th) 45.0 dBAE
Dynamic trigger boundary set above steady-state hydraulic flow noise to eliminate false triggers while maintaining high sensitivity to micro-cracks.
EDGE HARDWARE DSP Monpod FPGA processes waveforms at 2 MS/s to compute scalar features locally
ASTM E1316 & ISO 12716 Standardized

Delta-T Defect Triangulation

Because sound waves propagate through steel at a known velocity (~3,200 m/s for plate waves), a crack burst reaches nearby sensors at slightly different microsecond arrival times.

By calculating the difference in arrival times (Δt) across a cluster of three or four transducers positioned around a tank bottom or vessel circumference, Updem's localization algorithms calculate hyperbolic intersection coordinates.

The maintenance team receives an exact coordinate: "Defect cluster located at Tank 104, weld seam 4, height 1.2m, angle 134°." When your team does enter the vessel during turnaround, they walk straight to the defect without having to scan the entire 1,500 square meters of steel.

Interactive Delta-T Defect Triangulation

Click anywhere on the tank shell below to simulate an acoustic crack event and observe acoustic wave propagation.

Pinpoint Coordinates
X: 18.2m, Height: 6.1m
Delta-T Arrival Offsets
Δt(A-B): +14.2 µs
Recommended Reliability Action
Course #2 Shell Plate: Longitudinal weld stress concentration confirmed.

From Monpod™ Edge Sensors to Cloud Predictive Platform

Continuous acoustic emission monitoring produces massive data volumes: sampling at 1 MS/s generates gigabytes of raw data per hour. Storing raw acoustic waveforms in standard industrial databases causes massive network congestion and database bloat.

Updem solves this with a balanced two-tier architecture developed specifically for heavy industrial static infrastructure:

TIER 1: THE MONPOD™ AT THE EDGE

Autonomous Acoustic Capture & Feature Extraction

The Monpod™ is an autonomous, battery-powered wireless module that magnetically attaches to tanks, pipes, and vessel shells. It connects to dual differential piezoelectric transducers that listen in the 100 kHz to 1 MHz ultrasound band.

  • Zero Hot Work: High-holding neodymium magnetic feet clamp in under five minutes without shutdown.
  • Edge Extraction: Computes rolling physical features (MARSE energy, peak amplitude, duration, hit rate, RMS) right on the device.
  • Wireless MQTT Telemetry: Transmits clean low-rate scalar telemetry over LoRaWAN or cellular directly to the cloud.
TIER 2: UPDEM CLOUD PLATFORM

Predictive Maintenance & Explainable Alarms

The Updem Cloud Platform manages your asset hierarchy (mapping tanks, reactors, and pipeline segments) and provides continuous predictive intelligence.

  • Learned Baselines: Automatically characterizes normal acoustic behavior for each asset during steady-state operation.
  • Explainable Anomaly Detection: Detects physical feature drift before failure, stating exact causes: "AE energy and hit rate rose 4x above baseline over 6 hours."
  • Closed-Loop Feedback: Maintenance engineers confirm or dismiss alerts with a single click, turning field expertise into verified training labels.
UNIVERSAL PROTOCOL & SENSOR INTEROPERABILITY

Connect Any Sensor via MQTT, HTTPS, OPC UA, and Modbus

While our Monpod modules provide dedicated high-frequency acoustic monitoring for static equipment, the Updem Cloud Platform is protocol-agnostic. Plant operators can connect existing rotating machinery sensors, process instrumentation, and third-party transmitters into a single unified predictive maintenance interface:

MQTT & MQTTS
Lightweight, secure publish/subscribe telemetry for wireless sensor nodes, IoT gateways, and remote tank farms.
HTTPS & REST API
Standard REST endpoints and webhooks for cloud-to-cloud integration and batch synchronization with data lakes.
OPC UA & MODBUS TCP
Direct fieldbus connectivity to plant DCS racks, programmable logic controllers, and on-premise SCADA systems.
MULTI-SENSOR ECOSYSTEM
Correlate acoustic emissions with vibration sensors, temperature probes, line pressure, and ultrasonic thickness gauges.

Review Our Technical Whitepaper & Engineering Datasheet

Download our 18-page engineering brief: "Continuous Acoustic Emission Monitoring of API 650 Storage Tanks and Pressurized Static Equipment", or inspect our hardware specifications.

Request Technical Paper →
UPDEM OY
Industrial Acoustic Sensing Division • Tampere, Finland
DOCUMENT: DS-AE750P-REV4
CLASSIFICATION: TECHNICAL SPECIFICATION
DATE: SEPTEMBER 2026

Updem Monpod™ Wireless Acoustic Emission Module & Cloud Platform

Autonomous wireless acoustic emission hardware module with magnetic mounting feet and cloud predictive maintenance platform for continuous static asset integrity monitoring.

1. Monpod™ Wireless Hardware Specifications

Hardware System Monpod™ Wireless Module: Self-contained autonomous sensor enclosure with magnetic mounting feet and visual diagnostic LED.
Acoustic Transducers Dual differential piezoelectric ceramic transducers with low-noise balanced preamplification.
Acoustic Frequency Bandwidth 100 kHz to 1 MHz ultrasound passband (tuned to reject plant mechanical rumble).
Edge Feature Extraction Onboard real-time computation of MARSE energy, peak amplitude, duration, hit count, and RMS.
Dynamic Range 15 dBAE to 105 dBAE (Linear response > 90 dB).
Continuous Sampling Rate Up to 1 MS/s transient acoustic capture at the edge.
Mounting Method High-pull neodymium magnetic feet for 5-minute clamping onto steel without hot work or shutdowns.
High-Temperature Waveguides Direct surface up to +150°C; up to +550°C using solid 316L stainless acoustic waveguides.
Wireless Telemetry Uplink LoRaWAN and Cellular LTE-M / NB-IoT with MQTT protocol transmission.

2. Cloud Predictive Maintenance Platform Specifications

Cloud Infrastructure High-availability telemetry platform with centralized static asset hierarchy (tanks, reactors, pipelines).
Baseline Learning Engine Autonomous learned operational baseline per asset during steady-state production.
Anomaly Detection Unsupervised statistical and physical feature drift scoring before defect progression.
Alert Explainability Plain-language engineering directives detailing exact feature deviations (e.g. 4x energy rise over baseline).
Closed-Loop Feedback One-click engineer confirmation loop (confirm / false alarm) continuously training site models.
Standards & CMMS Links API 650/653 Annex C, API 581 RBI, ASME Section VIII, REST API, MQTT, and SAP PM / Maximo connectors.
Engineered by Updem Oy • Platform 6, Tampere, Finland • Tel: +358 46 847 2441